AI Assistant

The AI Assistant is a conversational tool developed by Magnific for creating and editing images. The AI Assistant is Magnific's conversational tool for creating and editing images. Describe what you want in a chat, and the Assistant generates or edits visuals in real time. You can also upload your own images, reuse previously generated ones, and apply styles, compositions, or guided workflows — all through the same interface.

How to access the AI Assistant

Go directly to magnific.com/app/assistant
From the Magnific homepage, click AI Assistant
From the AI Suite top menu, click Assistant under the Image section.

How does the AI Assistant work?

The AI Assistant uses a real-time chat interface. You type what you want — a new image, an edit to an existing one, or a variation — and the Assistant generates it. Every message builds on the previous context, so you can iterate without starting over. A typical workflow looks like this:

Open the AI Assistant
Type a description of what you want
For example, "A flat-lay photo of a coffee cup on a marble table, warm light, top-down view"
The Assistant generates the image
Continue the conversation to refine it
"Make the cup blue", "Add a croissant next to it", "Change the angle to 45 degrees".

Choosing a model

The AI Assistant supports multiple generation models. Each one has different strengths — some are better for photorealism, others for stylized or artistic output.

You can switch models from the settings panel in the chat interface. Available models include GPT-4o, Flux Kontext, Google Nano Banana, and others. We're always adding new models and updates to our tools, so check back in the AI Assistant settings to see the latest options.

Auto: Lets the system choose the most suitable model based on your prompt.
GPT Medium / GPT HQ: General-purpose models with increasing quality and detail; GPT HQ gives richer detail and resolution than standard GPT.
Flux Kontext [Pro] / [Max]: Creative engines focused on visual coherence and layout control; Max provides tighter structure and text handling.
Seedream 4/ Seedream 4 4K: Advanced multimodal image generation and editing designed for high resolution and multi-image consistency; produces detailed visuals up to 4K and supports complex scene composition
Google Nano Banana / Google Nano Banana 2: Based on Google’s Gemini image models (Nano Banana is Gemini 2.5 Flash Image, with Nano Banana 2 offering improved reasoning and prompt handling); known for fast generation, strong realism, and reliable multi-image edits.
Runway Gen4: A model tuned for creative storytelling with reference support, good for scenes with consistent visual themes.
Imagen 4: A high-capability model optimized for realistic and expressive text-to-image generation.
Mystic 2.5 Fluid: Balances artistic style with detail; ideal for dreamy or stylized visuals.
Ideogram: Focused on graphic design and typography, great for posters, cards, and layouts that require readable text.

How to change the model

Open the AI Assistant
Click the model selector in the chat interface
Choose the model that fits your goal

Tip: If you're not sure which model to use, start with the auto. You can always switch mid-conversation and re-generate.

Setting aspect ratio, style, and colors

Before generating an image, you can define its format and visual direction using three settings in the horizontal bar below the chat input:

Aspect ratio

Click the aspect ratio button and choose a preset (1:1, 16:9, 9:16, 4:3, etc.). This ensures your image fits the intended format — social posts, thumbnails, banners, etc.

Style

Click the style icon (star) to apply a visual tone to your generations. You can choose from predefined styles or select one of your own Custom Styles for a consistent look across images.

Color palette

Click the palette icon to pick or add a custom color palette. This is useful for keeping output consistent with your brand colors or a specific creative direction.

All three settings are optional. If you skip them, the Assistant will use default settings based on your prompt.

Uploading and editing images

You're not limited to generating from scratch. You can upload your own images or reuse AI-generated ones and edit them through the chat.

How to upload and edit

Click the + button or drag and drop an image into the chat.
Describe the edit you want
For example, "Remove the person in the background", "Change the shirt color to red", "Make this look like a watercolor painting".
The Assistant applies the edit and shows the result.
Continue iterating as needed.

You can also click on any previously generated image in the conversation to use it as the starting point for a new edit.

Templates

Templates are guided processes designed to help you complete specific design tasks more efficiently, without writing prompts. They're ideal when you want to save time or follow a more structured creation process.

How to use templates

Click Templates in the chat interface
Choose one based on your task
For example, "generate scene from character", "adjust camera angle", "create product close-ups").
Follow the steps — the Assistant will ask for inputs like images, descriptions, or preferences
Review the result and refine if needed

These templates are built for multiple design needs and developed with input from creative professionals. They're ideal when you want to save time or prefer a more structured process over freeform prompting.

Tips & best practices

Be specific in your prompts. The more detail you provide (lighting, angle, mood, objects), the closer the result will be to what you imagine.
Iterate in the conversation. Each message builds on the previous result. Ask for changes one at a time for more predictable results.
Try different models. If one model doesn't nail the style you want, switch to another and re-generate from the same prompt.
Upload reference images. Showing what you want is often faster and more precise than describing it.
Use styles and color palettes for consistency. If you're generating a series of images (e.g., for a campaign), set the style and colors once and keep them throughout the session.
Start with templates for common tasks. If your task matches an available template, start there — it's faster than building the prompt yourself.

Usage limits and rights

The AI Assistant is available to all registered account. Each generation uses credits, and the amount may vary depending on the model. You can find the most up-to-date information on credit usage in this article. You can generate and download as many images as you like. Image downloads do not count toward your daily Magnific asset download limit.

All generated content is subject to Magnific's Terms and Conditions for AI Products. You're welcome to use AI-generated images for personal or professional use as long as they don't infringe on third-party IP rights.